Using nltk with Python 3 (8)

Chapter 8: Analyzing Sentence Structure

>>> trees = parser.nbest_parse(sent)
Traceback (most recent call last):
  File "...\lib\site-packages\nltk\parse\chart.py", line 530, in _add_index
    raise ValueError('Bad restriction: %s' % key)
ValueError: Bad restriction: next

Weird. The second time the next method was renamed to __next__. I suspect 2to3 at work here, as it makes no sense at all in the context. Change line #166 in parse/chart.py to:

def next(self):

And line #288 to:

def next(self):

And line #1047 to:

lhs = left_edge.next()):

And line #1002 to:

left_edge.next() == right_edge.lhs()):

And line #1118 to:

next, index = edge.next(), edge.end()

And line #1098 to:

for prod in grammar.productions(lhs=edge.next()):

And line #1219 to:

lhs=left_edge.next()):

---

>>> nltk.corpus.sinica_treebank.parsed_sents()[3450].draw()
Traceback (most recent call last):
  File "...\lib\site-packages\nltk\tree.py", line 1337, in sinica_parse
    treebank_string = string.join(tokens)
AttributeError: 'module' object has no attribute 'join'

And again an error 2to3 should have corrected. Change line #1337 in tree.py:

treebank_string = ' '.join(tokens)

While here, change line #238, to take care of the next error:

if not filter or filter(self):

Done with chapter eight.